Research map / Bandits and online learning / Contextual bandits and Thompson sampling
Misspecified Gaussian Process Bandit Optimization
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use phased uncertainty sampling and elimination without knowing ε | GP bandit optimization · Misspecified kernelized bandits | cumulative regret bound: same regret guarantee without knowledge of ε | Helped |
| Enlarge GP-UCB confidence bounds using the known misspecification error ε | EC-GP-UCB · Misspecified kernelized bandits | cumulative regret: O\Big( B \sqrt{\gamma_T T} + \sqrt{(\ln (1/\delta) + \gamma_{T})\gamma_T T} + \epsilon T \sqrt{\gamma_T}\Big) | Helped |
| Use standard GP-UCB without an ε-dependent confidence enlargement when ε is unknown | GP-UCB · Misspecified kernelized bandits | cumulative regret | Hurt |
| Combine EC-GP-UCB candidates with regret bound balancing | Contextual GP bandits with stochastic action sets · Misspecified contextual kernelized bandits | contextual cumulative regret bound: \tilde{O}\Big((B\sqrt{\gamma_T} + \gamma_T) \sqrt{T} + \epsilon T \sqrt{\gamma_T} + (B\sqrt{\gamma_T} + \gamma_T)^2 \Big | Helped |
| Use phased uncertainty sampling with the linear kernel | Phased GP Uncertainty Sampling · Misspecified linear bandits | cumulative regret bound: \tilde{O}(\epsilon T \sqrt{d}) | Helped |
| Allow misspecification of size ε | Any algorithm in the general kernelized setting · Misspecified kernelized bandits | cumulative regret lower bound: \Omega(T\epsilon) | Hurt |
Closest papers by meaning
- Improved Regret Analysis in Gaussian Process Bandits: Optimality for Noiseless Reward, RKHS norm, and Non-Stationary Variance (ICML 2025 · similarity 0.92)
- Sharper Guarantees for Misspecified Kernelized Bandit Optimization (NeurIPS 2026 · similarity 0.92)
- Gaussian Process Upper Confidence Bound Achieves Nearly-Optimal Regret in Noise-Free Gaussian Process Bandits (NeurIPS 2025 · similarity 0.92)
- Improved Regret Analysis For Parallel Gaussian Process Bandit Optimization (NeurIPS 2026 · similarity 0.90)
- Improved Regret Bounds for Gaussian Process Upper Confidence Bound in Bayesian Optimization (NeurIPS 2025 · similarity 0.89)
- Random Exploration in Bayesian Optimization: Order-Optimal Regret and Computational Efficiency (ICML 2024 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.